TFI International: An $18B Take-Private LBO
A full leveraged-buyout model and one-page Investment Committee memo for a take-private of TFI International (NYSE/TSX: TFII), framed as a CDPQ + Brookfield consortium — a Montreal-domiciled national champion where the Quebec pension is the natural home-team anchor and founder-CEO Alain Bédard is a credible rollover partner. The deal is structured at a US$187.88 offer (22% premium to the US$154 unaffected), a US$15.5B equity purchase and US$18.4B enterprise value, or 12.7x forward / 13.9x LTM EBITDA. The central analytical stance is discipline: at a full 12.7x entry multiple you cannot underwrite a classic 25% private-equity outcome, and the memo refuses to pretend otherwise. Instead it builds an infrastructure-style compounder thesis fit for long-duration, lower-hurdle Canadian capital — buy a cash machine (US$784M trailing free cash flow) at the early turn of the freight cycle, sweep debt from 4.9x to ~1.1x net leverage, ride mid-cycle EBITDA normalization from ~$1.45B to ~$2.08B, and exit at roughly the entry multiple. The base case returns a 13.9% IRR and 1.91x MOIC, and a value-creation bridge decomposes it to 68% EBITDA growth, 35% deleveraging, and a ~4% multiple de-rate as drag — an operational return, not a financially engineered one. The freight-cycle read is deliberately two-sided: Class 8 truck orders running +125% year-to-date confirm the cycle is turning, but the memo flags that the surge is partly EPA-2027 pre-buy and replacement demand, so it underwrites mid-cycle rather than peak.
Pipeline Architecture
Primary sources (Q2 2026 results, funded-debt/EBITDA leverage, TTM free cash flow, Class 8 truck-order data) feed a structured, currency-tagged fact base. Entry structure and a conservative operating plan are set, then routed through an eight-sheet LBO where sources equal uses and the debt schedule sweeps free cash flow through the hold; returns and two sensitivity tables fall out of the model, every value is re-checked in an independent formula engine, and the one-page IC memo ties to the model before the PDF renders.
Methodology notes
Honest returns before a pretty number: the model was built to whatever the freight-cycle and free-cash-flow assumptions actually produce, not reverse-engineered to a target IRR. At a 12.7x forward entry multiple the base case lands at ~14%, which is precisely the point — a full-price take-private of a quality compounder is an infrastructure-style, deleveraging-driven return for long-duration capital (CDPQ, Brookfield), not a 5x-levered private-equity flip. The memo states where the deal stops working (above ~13x forward the IRR falls below the consortium hurdle) rather than hiding it.
Value creation decomposed and stress-tested: returns break down to roughly two-thirds EBITDA normalization (margin ramping 19.3% to 21.7% off a freight-recession trough) and one-third deleveraging (100% excess free-cash-flow sweep retires the Term Loan B in full, ~$4.1B of debt reduction flowing to equity), with a small conservative exit de-rate (12.5x vs 12.7x entry) as drag. Two sensitivity tables flex exit multiple against EBITDA and entry leverage against exit multiple, showing IRR spanning ~8% to ~21% with the exit multiple as the dominant swing factor.
A model that ties, verified independently: sources equal uses to the dollar; interest is charged on beginning-of-year debt balances so the free-cash-flow sweep and the debt schedule resolve without a circular reference; and rather than trusting the spreadsheet on faith, every headline output (entry EV, sponsor equity, exit equity, MOIC, IRR, exit net debt) was re-evaluated in an independent Python formula engine and matched the intended logic exactly before shipping.
Two-sided cyclical honesty: the Class 8 truck-order surge (+125% year-to-date, June +230%) is the freight-recovery evidence, but the memo explicitly discounts it as partly EPA-2027 NOx pre-buy and fleet replacement rather than durable end-demand, and warns that the same new capacity can cap 2027-2028 rates. The plan therefore underwrites mid-cycle EBITDA, and the first named risk carries a concrete kill metric (Truckload operating margin below ~11% for two consecutive quarters).
What this isn’t (yet)
The honest limits. A page called “honest” with no limitations would be a credibility own-goal.
Independent · not investment advice
This is an independent piece of student work. It was not commissioned by, paid for by, affiliated with, or endorsed by CDPQ, Brookfield, TFI International, or any institution. The consortium framing is an analytical construct; there is no actual or rumored transaction. The PROCEED recommendation and all returns are one student's model output on disclosed facts and public data. Nothing here is a recommendation.
Modeling simplifications carried visibly
Interest is charged on beginning-of-year balances to keep the debt sweep non-circular, a standard teaching simplification that slightly understates in-year interest. Entry net debt (~$2.9B) is derived from the most recent long-term-debt disclosure plus current maturities and cross-checked against the 2.4x funded-debt/EBITDA ratio, not lifted from a single line item. The operating plan is a clean five-line build (revenue, margin, D&A, capex, working capital) rather than a bottom-up segment forecast, and the sensitivity tables hold non-swung inputs at base.
A framing, not a live deal
The sponsor identities, offer premium, and financing mix are chosen to make a coherent and defensible case, not sourced from any transaction process. Real diligence would test the Truckload cost structure, contract renewal cadence, acquisition pipeline, and financing terms far more deeply than a one-page memo and an eight-sheet model can. The artifact proves LBO mechanics and investment judgment, not deal access.